Issue · 2026-06-09
Auto-triage customer emails with a 15-minute Zapier + GPT workflow
Most small service businesses lose leads in a shared inbox. Quote requests sit next to newsletters. Nobody owns triage, so first response time drifts to 24–48 hours — and the prospect hires someone who replied faster.
This playbook shows you how to auto-triage customer emails in about 15 minutes using Gmail, Zapier, and a simple GPT classification prompt. No developer, no enterprise helpdesk — just a workflow you can copy today.
What you'll automate
- New email hits a shared Gmail label (
inbox/shared). - A structured prompt classifies intent (quote, support, spam) and urgency (high/low).
- Urgent quote requests ping Slack (or your team channel). Support threads get a draft reply saved in Gmail. Spam gets archived.
Who this is for
- 1–10 person service businesses sharing one
hello@inbox - Owners who quote from email and can't afford a missed lead
- Teams already on Gmail + Zapier (free tier works for low volume)
What you get free on this page
The full step-by-step setup — trigger, classification prompt, routing rules, and what to measure — is published here. Want the Zapier export, GPT prompt file, and checklist in one download? See the Inbox Triage Kit.
Pairs well with: Quote follow-up automation — triage incoming requests first, then chase silent quotes automatically.
Frequently asked questions
How long does email triage automation take to set up?
Most teams finish the first version in about 15 minutes using Gmail and Zapier. Tune the classification prompt after 20 real emails.
Do I need to code?
No. Every step uses no-code tools with copy-paste templates. The GPT step runs inside Zapier — you paste a prompt, not Python.
Will this work with Outlook or Microsoft 365?
This playbook uses Gmail labels and Zapier's Gmail trigger. For Outlook, swap the trigger for Microsoft Outlook in Zapier — the classify-and-route logic stays the same.
What tools do I need?
Gmail (shared label or filter), Zapier (free tier OK for low volume), and optionally Slack for urgent alerts.
How do I know it's working?
Track median first-response time and classification accuracy. Target 80%+ correct intent labels before adding more automation.
Is the full playbook really free?
Yes — the complete setup guide is on this page. Optional paid kit adds export files and a printable checklist.
Most small shops lose leads in a shared inbox. This playbook routes urgent requests, drafts replies, and logs everything — without hiring ops.
What happens
The problem
When three people share one Gmail inbox, urgent quote requests sit next to newsletters and spam. Nobody owns triage, so response time drifts to 24–48 hours.
What you'll build
- Trigger — new email in a Gmail shared label (
inbox/shared) - Classify — structured GPT prompt returns intent + urgency
- Route — Slack alert for urgent items; draft reply saved as Gmail draft
Step-by-step
1. Create the Gmail label
Add a filter that applies label inbox/shared to messages sent to hello@yourbusiness.com.
2. Zapier trigger
New Email Matching Search in Gmail → search label:inbox/shared.
3. Classification prompt
Given this email, return JSON only:
{ "intent": "quote|support|spam", "urgency": "high|low", "summary": "one sentence" }
Email:
{{body}}
4. Route by urgency
urgency=high→ post to#sales-urgentin Slack with summaryintent=support→ create draft reply with your support macrointent=spam→ archive and labelauto-archived
Estimated setup time
15 minutes for v1. Tune the prompt after 20 real emails.
What to measure
Track median first-response time and % of emails auto-classified correctly. Target 80%+ accuracy before adding more automation.
Get the next playbook in your inbox — one practical automation every 7 days.
Tools for this playbook
Sign up through these links to implement the workflow. Replacing a tool? The steps above still apply — swap the platform in the trigger and routing steps.
- Zapier — No-code automation for Gmail triggers, routing, and drafts.
- Make.com — Visual alternative if you prefer Make over Zapier for the same workflow.
- OpenAI API — GPT classification step for intent and urgency labels.
Some links in this playbook are affiliate links. If you sign up through them, Automate This Week may earn a small commission at no extra cost to you. We only recommend tools we use in the playbooks. Including NovaRho affiliate links where noted.